What 'AI-operated MSP' actually means (and what it doesn't)
Every MSP claims to be 'modern' these days. Few define what changes. Here's our definition of AI-operated — and the parts that are still, deliberately, human.
AI-operated is the term we use to describe how Ultiblob runs the platform. It's not a chatbot. It's not a marketing checkbox. It's a structural choice about which decisions get made by agents and which stay with senior engineers.
Agents handle three categories of work: detection (parsing telemetry into incidents), remediation (executing known-good runbooks against known-good signatures), and reporting (drafting customer comms, tickets, compliance evidence).
Humans handle one category: trade-offs. Every change that touches a customer environment goes through a senior engineer, even when the agent proposed it. Every incident that escalates past the runbook gets a human in 4 minutes or less (we publish the SLA).
The split exists because agents are excellent at toil and senior engineers are excellent at judgment. The MSP-industry default is the reverse: senior engineers doing toil, junior engineers making trade-offs. AI-operated inverts it.
What this means for customers: tickets resolve faster, but the hard ones get a name and a face. Compliance evidence is generated automatically, but a human signs the report. Cost optimizations are surfaced weekly, but the customer chooses what to apply.
The shortcut question we get: 'how big is your NOC?' The honest answer: smaller than you'd expect, because agents cover the breadth and our senior engineers cover the depth. Quality of judgment per ticket is the metric that matters, and that's the one AI-operated optimizes.
